Digital Transformation to Stay Operational

Why do plants decide to transform digitally? How are plant operations changed thanks to digital transformation, and does it help during #COVD19? And what does the I&C department have to do to make it happen? Imagine the peace of mind knowing the plant is operational, particularly in your area of responsibility. But how do operational excellence objectives translate into underlying enabling technologies? Let’s dig into it, level by level, until we get to the bottom of this. Here are my personal thoughts:

Operational Excellence

Operational excellence objectives include lowering operational cost, increasing operational revenue, and reducing operational risk. This is achieved by operating more consistently and reliably through continuous improvement in all operational departments; operational efficiency. That is, the underlying goals for reliability are greater availability, lower maintenance cost, extended equipment life, greater integrity, shorter turnarounds, and longer between turnarounds. The goals for process energy are lower energy consumption and cost, plus reduced emissions and carbon footprint. The goals for occupational safety and health is fewer incidents, faster response time, and lower incident costs. The goals for production are reduced off-spec product (higher quality), greater throughput, greater flexibility for feedstock and product grades, reduced operations cost, and shorter lead-time. All departments see improved productivity, reduced risk, and compliance. Reduced consumption and release to the environment means greater sustainability.

Operational Reports

If you have been operating consistently and reliably is confirmed from operational results for every operational department. Traditionally periodic reports are provided to supervisors and managers; with yearly, quarterly, monthly, weekly, or daily reporting period. But such operational reports may not have been provided for all operational departments; perhaps only one for production and another for quality, but not for energy efficiency or reliability. The operational reports contain operational information including the Key Performance Indicators (KPI) for the respective department with comparison against the targets set for the period and year. The reports drive corrective action. Traditionally reports have been compiled manually, but manually compiling reports with data from multiple teams is time consuming. With digital transformation reports are automatically generated. A common requirement for a digital transformation project is to provide operational reports for all operational departments: reliability, maintenance, integrity, process/energy, HS&E, production, quality, and for the plant manager.

The operational reports are personalized. Meaning each person gets a different report and the information included is selected depending on the roles and responsibility of each person. The reliability manager’s report is different from the production manager’s report. And a reliability engineer’s dashboard is different from that of the reliability manager.

However, the most interesting fact is that KPIs and periodic reports are snapshots and tend to be lagging indicators; meaning they measure outcomes which only indicate if efforts during the past reporting period have been successful or not. KPIs are usually calculated on historical data collected over the reporting period. That is, KPIs and periodic report tend to be reactive rather than predictive. With lagging indicators it is harder to pinpoint when and what is going wrong as it happens; causing the KPI to not be met at the end of the reporting period. Don’t get me wrong, this after-the-fact feedback is required to drive long-term improvement, but plant personnel at an operational level also need real-time information to become more predictive and specific.

Note that the content in operational reports is different from content in business reports. Operational reports visualize operational information such as number of unhealthy and low performing equipment, energy consumption, flaring volume, production volume, off-spec volume, quality (yield), batch processing times and time between batches, consumables used, and number of pipe sections to be retired; extracted from the plant historian platform based on underlying predictive analytics and advanced sensors. Business reports on the other hand visualize business information such as revenue, costs, inventory, and man-hours extracted from the ERP.

Continuous improvement means making sure KPIs are met and exceeded and getting better over time. Also as part of continuous improvement, the KPI selection and report content is reviewed for relevance from time to time. A new KPI may be added and a KPI not helpful may be removed. This in turn impacts the underlying analytics and sensors the I&C department must deploy.

Operational Dashboards

Because KPIs tend to be lagging indicators, meeting these KPI goals in turn requires real-time operational visibility and response to everything that goes on in the plant, as it develops. This visibility is provided through dashboards with operational information including real-time operational indexes (indices) which are leading indicators; meaning they show the current rate of change and the underlying conditions to make operations right, such as fouling, wear, production flow (rate), setpoint deviations, control loops in manual, and pipe wall thinning. Those are factors that lead to failure, inefficiency, missed production targets, off-spec product, cleanups, over-consumption, and loss of containment. That is, indexes that make work more predictive. Dashboards drive proactive work. A common requirement for a digital transformation project is to provide operational dashboards for all operational personnel: reliability, maintenance, integrity, process/energy, HS&E, production, quality, and for the plant manager.

The operational dashboards are personalized. Meaning everybody has a different dashboard and the information displayed is selected depending on the roles and responsibility of each person. The maintenance manager’s dashboard is different from the safety manager’s dashboard. And a maintenance engineer’s dashboard is different from that of the maintenance manager.

Note: real-time indexes are different from KPIs in that real-time indexes are leading indicators and KPIs tend to be lagging indicators. KPIs tend to be calculated on historical data while real-time indexes are calculated on live data. Also note that the content in operational dashboards is different from content in business management dashboards. Business dashboards cover financial, customer relationship management (CRM), and supply chain management (SCM).

The dashboards are web-based accessible everywhere as needed. It may be displayed on a large wall screen in the control room, or displayed on a desktop screen or laptop at your desk in the admin building, in the boardroom, or on-the-go from a tablet or smartphone. That is, you could be part of the team that work from home during COVID-19 or similar special circumstances with an at-a-glance view of your area of responsibility.

As part of continuous improvement, the operational index selection and dashboard content is reviewed for relevance from time to time. A new index may be added and an index not helpful may be removed. This in turn impacts the underlying analytics and sensors the I&C department must deploy.

Operational Notifications

Plant personnel are not looking at their dashboards all the time. Therefore they need operational notifications when something is going on in their respective area of responsibility. That is, they receive a message on their smartphone when some software or subsystem in the plant predicts a problem related to them. That is, notifications essentially are alarms relayed outside the system to wherever you are. A common requirement for a digital transformation project is to provide operational notifications for all operational personnel: reliability, maintenance, integrity, process/energy, HS&E, production, quality, and for the plant manager.

The operational notifications are personalized. Meaning the notification sent and the information in the nonfiction is selected depending on the roles and responsibility of each person. The integrity manager’s notifications are not the same as those received by the quality manager. And an integrity engineer’s notifications are not the same as those received by the integrity manager.

The idea is not to create as many kinds of notifications as possible, because that will simply flood the recipient and lead to frustration. Instead create ‘good’ notifications which are relevant to the recipient's priorities, unique and not merely a repetition of information from another notification, timely meaning neither long before intervention is necessary nor too late for action to be taken, prioritized flagging the urgency of the issue predicted or detected, a clear understandable message that is easily understood, diagnostic describing the problem, and advisory recommending action.

Since the notifications go to your phone, they reach you wherever you might be: in the control room, at your desk in the admin building, in the boardroom, or even outside the plant. That is, you could be part of the team that work from home during COVID-19 or similar special circumstances and still receive notifications as needed.

As part of continuous improvement, the operational notification selection and content is reviewed for relevance from time to time. A new notification may be created and a notification not helpful may be removed. This in turn impacts the underlying analytics and sensors the I&C department must deploy.

Operational Information

Operational reports, dashboards, and notifications contain operational information such as KPIs, real-time indexes, and alarms required by the various operational departments in the plant: reliability, maintenance, integrity, process/energy, HS&E, production, quality, and for the plant manager. This information is usually not raw data direct from a sensor. Rather, information is distilled from data from multiple underlying sources. The raw data originates from permanent sensors, portable testers, or manual data entry. KPIs and real-time indexes are typically calculated from multiple data points. These data points in turn may be a result from operational analytics such as predicting process values or from totalization.

Note that operational information is different from business information. Business information such as revenue, costs, inventory, and man-hours reside in the ERP, CRM, and SCM systems. This data is transactional rather than real-time. Operational information such as process, equipment condition, and energy consumption etc. is streaming live in real-time directly from sensors and automation systems or through the historian platform.

As part of continuous improvement, the operational information in reports, dashboards, and notifications are reviewed for relevance from time to time. New information may be included and information not required may be removed. This in turn impacts the underlying analytics and sensors the I&C department must deploy.

Operational Data

Operational data comes from the various automation systems in the plant such as RTU, PLC, SIS, DCS, Machinery Protection Systems (MPS), power meters, weighing scales, inventory tank gauging systems, vibration monitoring systems, Intelligent Device Management (IDM) systems, wireless sensor networks (WSN), and other subsystems. These systems in turn get their data from sensors, portable testers, or in some cases from manual data entry. Manually collecting data using portable testers to surveillance/inspection rounds manually noting down readings on clipboards or even tablets is not effective. It is very time consuming, yet the data is too infrequent to be predictive. A basic requirement for all digital transformation projects is digitization of data collection from the plant, manual data entry, and manual data copying from one system to the other.

Plants already have sensors, lots of sensors, but they are mostly process sensors for Core Process Control (CPC) which is nearly fully automated in the DCS. However, most maintenance, reliability, integrity, process energy, and Occupational Safety and Health (OSH) monitoring today is manual. Digital transformation of these tasks require lots of additional advanced sensors for condition monitoring, efficiency monitoring, and situational awareness. Wiring sensors point-to-point in existing “brownfield” plants is not practical. Therefore many of these additional sensors will be wireless. Wireless sensors for these functions include corrosion, erosion, vibration, position, level, level switch, contact, hydrocarbon or chemical leak, flow, pressure, temperature, acoustic noise, H2S gas, CO gas, Oxygen depletion, and more. These are advanced industrial sensors available from your automation suppliers. Often they are wireless versions of the same sensors used for core process control. This means, the portion of data stored in the historian platform used for Monitoring and Optimization (M+O) will be much larger than it is today. In the future, both CPC and M+O sensors may be networked using Ethernet-APL.

This also means the portion of sensors used for Monitoring and Optimization (M+O) will be much larger than it is today. Additionally it means the portion of sensors which are wireless will be much larger too.

Automating the data collection with sensors reduces the number of staff and external contractors that need to walk about the plant to collect data. Personnel can instead focus on acting on the data. Not only does automatic data collection make the plant more productive and predictive, but during COVID-19 or similar special circumstances it enables the plant to operate with reduced personnel at site, and a greater portion of staff working from home.

Engineers have an innate ability to figure out what sensors are required to address a certain problem. The additional sensors for M+O functions like condition, efficiency, and safety as well as the supporting Wireless Sensor Network (WSN) are deployed and maintained by the I&C department, just like the sensors on the DCS for CPC, also integrate them with the historian platform and other software. This typically happens gradually; starting in one plant unit, expanding throughout the area, and then plant-wide. It may start with sensors for energy efficiency such as steam trap or PRV monitoring, or it may start with condition monitoring, and then expand to include sensors supporting the needs of the other operational departments: reliability, maintenance, integrity, process/energy, HS&E, production, and quality. Step by step the plant is fully instrumented.

Operational Analytics

Manually interpreting vibration data, pipe wall thickness data, acoustic noise, and other data is very time consuming. Due to lack of resources, data for each piece of equipment may only be looked at maybe once a month or once a year. This is not predictive. A common requirement for a digital transformation project is therefore to automatically predict failure, fouling, and process upsets. Operational analytics interpret operational data aggregated from multiple sources such as from sensors, portable testers, and other systems such as the historian platform, distilling it into operational information. That is, some of this data comes from sensors wired to the control system, other data may come direct from wireless sensors. For instance, sensors installed on equipment pick up on early symptoms of developing problems. Operational analytics such as condition monitoring apps in turn interpret the raw sensor data to predict problems and diagnose the problem, to provide a description of the problem: descriptive analytics. For simplicity there are purpose-built apps, readymade for pumps, compressors, cooling towers, and more. For each failure mode there are recommended actions prescribed; prescriptive analytics. Similarly, there are performance analytics apps for efficiency monitoring that use data from multiple sensors to compute the efficiency of equipment like heat exchangers, compressors, and others to predict fouling to optimize cleaning. A third example is corrosion analytics that use data from corrosion sensors to predict pipe section retirement, optimize crude blend, and corrosion inhibitor injection. Lastly, process analytics predicts process upsets to give operators early warning to respond. These analytics software apps are available from automation vendors.

The information coming out of the analytics, such as pending equipment failure or process upset prediction, efficiency value, or inferential measurement can be seen in the analytics app itself, but also goes into reports and dashboards, and also reach the responsible persons as notifications when set limits are exceeded. That is, you could be part of the team that work from home during COVID-19 or similar special circumstances and still receive notifications as needed.

Note that operational analytics is different from business analytics. Operational analytics uses live-streaming process data from sensors. The analytics models are mostly readymade first principles (1P) and FMEA fault trees for the various equipment types found in a plant. This enables I&C engineers with little or no data science skills to deploy analytics in a plant. But some “use cases” where 1P and FMEA is not established need some type of AI machine learning algorithm to uncover such knowledge and build the predictive model. Operational analytics apps have OPC-UA interface to easily connect with process automation systems and the underlying sensors. Business analytics on the other hand extracts transactional data from the ERP to apply analytics on this mostly historical data, and visualize it graphically as Business Intelligence (BI).

The additional analytics apps for M+O functions like condition, efficiency, and safety as well as the underlying application framework including server and VM are deployed and maintained by the I&C department, just like the apps in the DCS for CPC. This typically happens gradually; starting with one use-case and then the next. It may start with apps for steam trap or PRV monitoring, or it may start with an app for pump or compressor condition monitoring, and then expand to include apps supporting the needs of the other operational departments: reliability, maintenance, integrity, process/energy, HS&E, production, and quality.

Operational Software

Operational software for digital transformation includes not only the historian platform and various kinds of analytics apps, but also various kinds of software for presentation of information to plant personnel such as the dashboards, notifications, and Augmented Reality (AR), as well as mobility, simulation like digital twin and Virtual Reality (VR), IT/OT integration with ERP. Plants do not deploy software from all these categories at the same time. Rather apps are rolled out over time to gradually be absorbed into daily work practices. Mobility software is another useful component to enable some plant personnel to work from home during COVID-19 or similar special circumstances and still be able to support the plant.

In the past many apps used to run stand-alone on a laptop used only by a single engineer which was not good. Stand-alone is still possible, but by instead installing that software on a server on the network, it can be used by many persons, even at the same time. This increases the adoption of software-based and data-driven ways of working. There are specialized apps for each person’s task: reliability, maintenance, integrity, process/energy, HS&E, production, quality, and for the plant manager. This software is available from automation vendors and need not be custom made. Note that these apps don’t have to be installed on the same server or sit on the same ‘single platform’ to enable people throughout the plant to access them. That type of centralized monolithic architecture is reminiscent of the 1980s. Instead it is more practical to have apps used for different functions on separate servers, but it can be virtual machines (VM) on a common physical server. This modular architecture easier to gradually deploy and maintain.

Note that operational software is different from business software. Operational software apps integrate with each other using the industry standard OPC-UA and the earlier OPC Classic (DA, A&E, and HDA) which does not require any programming, just simple configuration. This enables I&C engineers without programming skills to deploy operational software. Business software apps on the other hand integrate with each other using non-standard proprietary (owned and licensed by a single vendor) API requiring custom programming.

The additional software apps for M+O are deployed and maintained by the I&C department, just like the apps in the DCS for CPC.

Operational Infrastructure

Digital transformation to achieve operational excellence requires a lot of additional automation components for the M+O functionality. This includes many software apps, many networks, and many advanced sensors. Collectively this additional automation is referred to as the Digital Operational Infrastructure (DOI). In order to not interfere with the robustness and safety of the DCS performing the core process control, the DOI is deployed as a second layer of automation on the side of the DCS as per the NAMUR Open Architecture (NOA) for Industrie 4.0 (Industry 4.0).  A common requirement for a digital transformation project is to provide DOI that does not interfere with the existing automation in the plant.

(Pyramid courtesy NAMUR)

The software apps and sensors include all those discussed above. Networks include the Wireless Sensor Network (WSN) discussed above, as well as industrial Wi-Fi to enable use of tablets, wearable tablets, and phones in the plant. Additionally there is Ethernet at level 3 Central M+O and level 2 plant M+O. OPC-UA is the core data transport between applications at these two levels. This networking is available from automation vendors. In the future there will also be 2-wire Ethernet-APL for sensors and valves at level 1. Level 3 software for M+O can be made accessible from outside the plant to enable some plant personnel to work from home during COVID-19 or similar special circumstances and still be able to support the plant.

Note that the DOI is different from office infrastructure for business systems like the ERP and BI. The analytics and other software are different as explained above. Both the DOI and the admin office have Ethernet and Wi-Fi, but the application protocols running over these networks are different. The DOI for M+O is deployed and maintained by the I&C department, just like the DCS, sensors, valves, and other systems for CPC to support the other operational departments: reliability, maintenance, integrity, process/energy, HS&E, production, and quality.

Operational Technology

The DCS, and all the existing automation systems around it, as well as the new DOI, are now often referred to collectively as Operational Technology. In plants it is better known as Instrumentation and Controls (I&C) or Industrial Control System (ICS).

Level 3 higher-level analytics software predicts problems in plant equipment, but the CMMS used to manage work orders sit at the level 4 ERP system. Manually creating work request tickets in the ERP/CMMS whenever some type of problem is predicted would be very time consuming. Therefore a long-term requirement for a digital transformation project is to integrate operational analytics with the ERP/CMMS to create a digital workflow. Since level 0-3 is OT, and level 4 is IT, this part of a DX project requires IT/OT integration and IT/OT collaboration. Similarly, your automation vendor collaborates with your IT vendor for the solution.

IT also provides the communication path from the DOI to the Internet. Again, IT/OT collaborate on this internet connection. This makes the level 3 software tools accessible from outside the plant to enable some plant personnel to work from home during COVID-19 or similar special circumstances and still be able to support the plant.

Note that operational technology (OT) is different from office Information Technology (IT). Thanks to OPC-UA, and OPC Classic before it, a DCS can be connected to a third-party historian platform with dashboards and notifications, high-level condition and performance analytics, detail vibration, corrosion, and valve analytics, intelligent device management, and AR etc. all from other vendors. On the other hand, an ERP system is totally monolithic.

The operational technology (OT) including the DOI and DCS are deployed and maintained by the I&C department to support the other operational departments: reliability, maintenance, integrity, process/energy, HS&E, production, and quality.

Operational Readiness

The “end state” of the digitalization journey is to build out the operational technology with digital operational infrastructure including operational software such as operational analytics that distills new operational data from additional sensors into operational information which goes into operational notifications, operational dashboards, and operational reports which enables plant personnel to change from manual and paper-based ways of working to new digital, automatic, software-based, and data-driven ways of working to improve reliability, maintenance, integrity, energy efficiency, HS&E, production, and quality – which is operational excellence. Having these modern digital software tools in place also helps attracting and retaining the right people in the plant. The DOI is also an important part of the plant’s business continuity plan; operational readiness to enable some plant personnel to work from home during COVID-19 or similar special circumstances and still be able to support the plant.

With COVID-19 the need for a Fourth Industrial Revolution (4IR) and operational technology has increased. Schedule a meeting (online) with the operational departments and the I&C team to plan for digital transformation. Forward this essay to your manager and colleagues. And remember, as there is a lot of hype around platforms and analytics, always ask for product data sheet to make sure the product is proven, and pay close attention to software screen captures in it to see if it does what is promised without expensive customization. Well, that’s my personal opinion. If you are interested in digital transformation in the process industries click “Follow” by my photo to not miss future updates. Click “Like” if you found this useful to you and “Share” it with others if you think it would be useful to them.